[ADD] Kaggle submission

configs/kaggle_submission.yml

@@ -0,0 +1,13 @@
+command:
+  (): geoguessr_bot.commands.KaggleSubmissionCommand
+  guessr:
+    (): geoguessr_bot.guessr.DinoV2Embedder
+    embedder:
+      (): geoguessr_bot.retriever.DinoV2Embedder
+      device: "cpu"
+    retriever:
+      (): geoguessr_bot.retriever.Retriever
+      embeddings_path: !path "../resources/embeddings.npy"
+    metadata_path: !path "../resources/metadata3.csv"
+  image_folder_path: !path "../data/kaggle/images"
+  output_path: !path "../data/kaggle/submission.csv"

geoguessr_bot/commands/__init__.py

@@ -1,2 +1,3 @@
 from .abstract_command import AbstractCommand
 from .embed_command import EmbedCommand
+from .kaggle_submission_command import KaggleSubmissionCommand

geoguessr_bot/commands/kaggle_submission_command.py

@@ -0,0 +1,30 @@
+import os
+from dataclasses import dataclass
+
+import pandas as pd
+
+from geoguessr_bot.commands import AbstractCommand
+from geoguessr_bot.guessr import AbstractGuessr
+
+
+@dataclass
+class KaggleSubmissionCommand(AbstractCommand):
+    """Submit a prediction to Kaggle
+    """
+    image_folder_path: str
+    output_path: str
+    guessr: AbstractGuessr
+
+    def run(self) -> None:
+        images_ids, latitudes, longitudes = [], [], []
+        for image_name in os.listdir(self.image_folder_path):
+            image_path = os.path.join(self.image_folder_path, image_name)
+            coordinate = self.guessr.guess_from_path(image_path)
+            images_ids.append(image_name.split(".")[0])
+            latitudes.append(coordinate.latitude)
+            longitudes.append(coordinate.longitude)
+        pd.DataFrame(dict(
+            image_id=images_ids,
+            latitude=latitudes,
+            longitude=longitudes,
+        )).to_csv(self.output_path, index=False)